Pi Compute Server for http://www.reddit.com/r/dailyprogrammer/comments/1qply1/111513_challenge_129_hard_baking_pi/

gistfile1.cpp

#include <cmath>
#include <cstdint>
#include <cstdio>
#include <cstdlib>
#include <iostream>
#include <limits>
#include <string>
#include <vector>

#include "Poco/Net/ServerSocket.h"
#include "Poco/Net/SocketAddress.h"
#include "Poco/Net/StreamSocket.h"
#include "Poco/Timespan.h"

using namespace std;

int getNextDigit() {
  static int next_digit = -1;
  ++next_digit;
  return next_digit;
}

int main() {
  Poco::Net::ServerSocket srv(9090);
  cout << "Waiting for clients ..." << endl;

  // Arbitrary number of clients are supported. Each client is given the 'next'
  // digit once it's done.

  double mysum = 3.0;  // used to sanity check
  static const float kComparisonMultiplier = 10000.0;
  for (;;) {
    Poco::Net::StreamSocket ss = srv.acceptConnection();
    ss.setReceiveTimeout(20000);
    ss.setSendTimeout(20000);
    int last_digit;
    unsigned int computed_value;
    char client_buf[10];
    ss.receiveBytes(client_buf, 10);
    sscanf(client_buf, "%d-%u", &last_digit, &computed_value);
    cout << "Received " << last_digit
         << "th digit as " << hex << computed_value << dec << endl;
    if (computed_value != 1000) {
      mysum += (1.0 * computed_value / pow(16.0, (last_digit + 1)));
      cout << "  Divergence from PI "
           << "(to " << kComparisonMultiplier << " accuracy) so far = "
           << fabs((mysum - M_PI) * kComparisonMultiplier)  << endl;
    }

    // Send next command
    int next_digit = getNextDigit();
    snprintf(client_buf, 10, "*%d*", next_digit);
    ss.sendBytes(client_buf, 10);
    cout << "Sent " << client_buf << endl;
  }
  return 0;
}

/*
 * Input Description
 * There is no formal input description, though this is the desired behavior:
 * You launch your main dispatcher-application on Computer A. You then launch
 * the computing-applications on Computer W, X, Y and Z. Computer A wants to
 * compute the first four digits of Pi, and sends the appropriate network
 * commands, one to each computer. Computer Y returns a result first, so the
 * dispatcher receives the data, saves in your output file the line "0:2", and
 * then gives the command to compute the 5th digit of Pi. Computers X, Z, W
 * finish in that order, returning the results to the dispatcher, which in turn
 * saves in the same format. They are then asked to compute digits 6, 7, and 8.
 * This repeats until your dispatcher application sends a "stop" or "kill"
 * command to the computing processes. It is up to you how many hexadecimal
 * digits each process computes.
 * Output Description
 * For each computed base-16 (hexadecimal) digit of Pi, write to a file a line
 * of text in the format of <Digit-Index>:<Computed-Digit>. The order does not
 * matter, and you may skip digits. An example of the file, after eight computed
 * digits, would be as follows:
 * 0:2
 * 1:4
 * 2:3
 * 3:F
 * 4:6
 * 5:A
 * 6:8
 * 7:8
 */